Public Discussions > General Discussion

What is the agenda for OS/2 today?

(1/4) > >>

Eugene Gorbunoff:
Hello Martin

What is the agenda for OS/2 today?
IMHO, open-sourcing is the direction, we all agree that the OS should include more open-source components.
But it's not agenda of OS/2.

Agenda for today:
What applications to develop to attract new users, extend the market?
What parts of the OS to improve? to modernize the system.

Martin Iturbide:
Hi Eugene.

I splited your topic, because the MenuetOS post was to show how other open source OSes are showing up.

About the agenda.  My suggestion had been posted on this blog since March of this year.
Which is start from "Top to down" on replacing the close source WPS Classes first and moving down (SOM, PM, the rest).

What I can recommend you is to start making your eCo Software open source and really start summing up to have a complete open source operating system in the long term. Choose the open source license you like and start opening the source.

If you need help choosing an open source license I can give you some advice. I'm not a lawyer but I know the basic stuff of how the most common open source licenses works.


Dear all,

open source has its virtues (anyone can develop it as the underlying code is freely available etc.), and the success of Linux reminds us that open source is powerful.  However, we must not forget that Linux was the first major open source operating system venture born in a time of increasing "anger" over the way MicroSoft operated and Windows worked (or did not work - sorry, but I could not help it).  The irritation over MicroSoft and Windows is still there, but to a lesser degree as Linux serves as a safety valve for those most disgruntled, and Windows 7 actually works quite well (though it still has its quirks).  Just like OS/2 - eComStation users like the OS/2 feel, Linux users have gotten to used to the Linux feel(s) [recall there are multiple Linux versions out there], and they are hence quite loyal.

But the success of Linux is also a result of customers paying for packaging (RedHat, Suse, etc.): this provides developers with funds to go further.  I think the future of OS2/2 - eComStation lies in being able to develop and maintain an operating system that sufficiently many people are willing to pay for.  Due to historical reasons, OS/2 - eComStation is unlikely to become fully open source as part of the underlying code is proprietary with a mix of IBM and MicroSoft ownership.  I am uncertain if this ownership/those patents are soon to expire (the 20 year lifetime of patents).  If that is the case, we are in a better situation to attract new developers, but also developers need funding.  In Linux that is taken care of by users paying for the packaging.  A bit like what it appears eco Software seeks to do.

But OS/2-eComStation will struggle to attract as many users as Linux has, just because they were the first as an open source operating system and therefore occupy this space.  To support this claim, look at Mozilla - they were the first and there are few other successful web browsers/email clients around.  In other parts of the computer world Open Office has been successful (the first open source office suite), and R is growing as a the first open source statistical analysis software covering most branches of statistics (where two of the strong points are much faster inclusion of new statistical procedures, and a fantastic open support community for users asking questions than any commercial statistical software can provide).  Mozilla, Open Office and R are all successes because they were first, and hence got the disgruntled commercial software users and attracted developers.  They also make it more difficult for other open initiatives in their part of the market because they've already captured these users and developers.

What does this mean for OS/2 and eComStation?

* It will be very difficult to succeed only sticking to the open source modus operandi, exemplified by OS/2 clones struggling even more than we are.
* Attempts to make OS/2-clones actually makes it harder to succeed for OS/2-eComStation as it spreads developer resources and reduce the the number of users (which again makes it less attractive to develop software and drivers).
In brief, the future for OS/2-eComStation lies in developing something that is so useful that people want to pay for it.  This development could be helped if the underlying code becomes open source once/if patent rights expire (and I believe they do in 3-4 years as there was little development in the source past OS/2 ver 4.52) as that makes life easier and more fun for developers.  But starting from scratch as Martin suggests is not a good idea as it spreads users who like the OS/2 feel on several operating systems.  While I am displeased with the many of the way Mensys works, we should acknowledge that without Mensys OS/2 would have died a long time ago.  But unless Mensys starts to deliver more and better, the risk is large that OS/2 is dead before patents expire for the important parts of the OS/2 source code.


Doug Bissett:

--- Quote ---This development could be helped if the underlying code becomes open source once/if patent rights expire (and I believe they do in 3-4 years as there was little development in the source past OS/2 ver 4.52) as that makes life easier and more fun for developers.
--- End quote ---

It will, IF, and only IF, the source code is made available. That could be the reason why IBM hinted that they have "lost" the source code. Once patent rights run out, nothing changes, unless the source code is made available, and I expect that that will never happen.

What it might do, is encourage more people to start deconstructing OS/2, as some of the Russians have been doing. However, I think that Mensys has been doing some of that too, to get to the most serious problems. That was how they came up with those extra fixes that were included with ACPI. As I understand it, they were told, by IBM, to do whatever they thought was necessary.

My view is, that anything that is developed from scratch, needs to be open source. But the WPS is definitely NOT the place to start, for a rebuild project. There are many smaller parts of OS/2 that basically stand alone, that could be reworked, with fewer people required, and a much better chance to show some progress in the near future. Some of that stuff has been started, by porting parts of Linux, but there has been little interest, so far, partly because the parts that have been done (BIND, DHCP) are not understood by most users, and many are afraid to try that stuff, simply because they are afraid to render their systems unusable. Also, there has been little "advertizing" about those things, so many people don't even know that they exist. OS2world does post announcements, which helps, but people don't bother looking at stuff that they don't understand.

One of the more troublesome parts of OS/2 is XCOPY (which may be one of the things that Microsoft still has some code in). The function of XCOPY is pretty well understood, and many people use it, but it is known that XCOPY has trouble with many thousands of files. Trying to fix the original program doesn't make sense, but it should not be too difficult for somebody to map out what it is supposed to do, and write a new program to do it. While they are at it, they might as well use resources that are available, like large memory, and AHCI function, that was not available when the original program was written.

MiniLVM is required to set up disk drives, but it also has some limitations that have not been resolved. Mensys has the code for that program, but I suspect that the faults are in the base LVM code. This will need to be fixed, if we ever hope to support large (>2TB) disk drives. I think that Mensys has done a preliminary evaluation, but I don't think much else has been done.

Then, there is the question of NIC drivers. Mensys has developed the NVETH (NVIDIA), Intel, and Realtek, NIC drivers (wired versions only, so far). They were always supposed to be developed as a template, so that other people could continue the project, and develop more drivers for more devices (RALINK, BROADCOM etc.). AFAIK, nobody has stepped up to do that. The same will be true after Mensys gets a template for wireless devices. There will also be requirements to update the current MultiMac drivers, as newer devices become available. Somebody had better become interested, or we are going nowhere fast.

USB devices: Wim Brule has stepped up to make video cameras work, but I think he is working alone, and I have no idea what license model he might be using (if any). He also did the Canon 4400F (?) scanner thing, and the USBECD driver, which makes it possible to interface with "unknown" USB devices (among other things).

Mensys does, of course, have other things that they are working on (ACPI, AHCI, etc.), but that is only a small part of what needs to be done, in the short term. UEFI is already causing a lot of trouble, with no resolution in sight.

OS/2 today, is so far behind, that it will likely be impossible to catch up. Many, talented, people are working at keeping it going, but little work is being done to make advances. It would take many more talented people to make progress. As noted, if somebody pulls those talented people out of the mix, to do something like rewrite the WPS, there will be nobody left to keep it going in the short term. If the base OS stops working, any work done on the WPS project will be for nothing. Of course, there are still virtual machines, and old hardware, to run it on, but that certainly isn't going to attract serious new users.

So, back to:
--- Quote ---What is the agenda for OS/2 today?
--- End quote ---

The way I see it, is that the agenda is to work as hard as possible, to extend the life as much as possible, with the resources that we have available. To do that, every OS/2 (eCS) user will need to contribute in one way, or another, and priorities will have to be established. Mensys has established some priorities, but they are subject to the reality that they need to be able to pay for what they do (including pay salaries). It is okay to dream about open sourcing the WPS, but doing it will mean that other areas will suffer. You can have the best WPS possible, but what good is it if you can't even boot OS/2, because it can't figure out UEFI? Those who have avoided upgrading to eCS, for one reason or another, are not helping the cause.

Second agenda: Somebody needs to win a BIG lottery, and donate a good chunk of it to new development. There was a case, recently, where some guy got whacked twice by lightening, in the same storm, and lived, so it may not be impossible.

Dear all,

thanks to Doug for extending the analysis I started on.  There are some common threads in my posting and Doug's follow up:

* Scarcity of resources - both funds and developers - that makes it imperative to have a strong focus.  In brief, while some things are nice to have, there are some bare necessities we must have.  This implies use the existing platform where it works as it is resource demanding to develop new stuff, and those resources have an opportunity value.
* When developing new stuff from the ground that is essential for OS/2-eComStation to work, seek open source to attract new developers.  It is important to note that the code can be open source, while users still pay for using the product.  While unpopular among some users, attracting revenues is necessary to cover development costs
* There is uncertainty about the source code ever becoming available.  If this is due to patent issues, it will eventually become available (I think in 2016).  If this is because IBM may have lost the source code (Doug's angle), it is more troublesome, but it would be a major blow for IBM if a software company looses code, someone in IBM will find it if the costs of admitting the code is lost is made sufficiently large.  To sum up: I think this uncertainty will be resolved in a way that spells a future for OS/2-eComStation.There are also large challenges caused by hardware development and user demands changing.  On the latter, smartphones and notepads make it even more important that communication works between PCs and these new devices work, and that wired and wireless net access.  On wireless, we also learned an open source lesson: we did not secure that the initial code was made available for new developers to continue the good work that initially was done.

Both US and EU competition legislation is quite tough in the sense that hardware developers cannot favor certain software developers.  This implies that hardware developers need to document what they do.  Here, I think that the OS/2-eComStation community can forge fruitful alliances with Linux and other developing environments to improve documentation.

Doug's closing wish is that someone needs to "win big" in a lottery to make more funds available.  Betting on Lady Luck is indeed risky business.  However, if we succeed in resolving the main eComStation challenges, we are more likely to attract investors or donors.  In my book these challenges are:

* Proper operation on modern hardware (at least for some producers).  Here, ACPI is key, and good work is done
* Connectivity to other devices (smart phones, notepads, existing and new wireless standards).  Here, much work remains.
* Key software that "all" users need: office suites, email clients, browsers, flash, and JAVA.  Here, good work is doneSo, the future is not that bleak.  But the eComStation community needs to prioritize, and here clearly spelled priorities by Mensys are part of the picture given their central role in the development of eComStation.



[0] Message Index

[#] Next page

Go to full version